Carving executive control at its joints: Working memory capacity predicts stimulus-stimulus, but not stimulus-response, conflict.

Three experiments examined the relation between working memory capacity (WMC) and 2 different forms of cognitive conflict: stimulus-stimulus (S-S) and stimulus-response (S-R) interference. Our goal was to test whether WMC's relation to conflict-task performance is mediated by stimulus-identification processes (captured by S-S conflict), response-selection processes (captured by S-R conflict), or both. In Experiment 1, subjects completed a single task presenting both S-S and S-R conflict trials, plus trials that combined the 2 conflict types. We limited ostensible goal-maintenance contributions to performance by requiring the same goal for all trial types and by presenting frequent conflict trials that reinforced the goal. WMC predicted resolution of S-S conflict as expected: Higher WMC subjects showed reduced response time interference. Although WMC also predicted S-R interference, here, higher WMC subjects showed increased error interference. Experiment 2A replicated these results in a version of the conflict task without combined S-S/S-R trials. Experiment 2B increased the proportion of congruent (nonconflict) trials to promote reliance on goal-maintenance processes. Here, higher WMC subjects resolved both S-S and S-R conflict more successfully than did lower WMC subjects. The results were consistent with Kane and Engle's (2003) 2-factor theory of cognitive control, according to which WMC predicts executive-task performance through goal-maintenance and conflict-resolution processes. However, the present results add specificity to the account by suggesting that higher WMC subjects better resolve cognitive conflict because they more efficiently select relevant stimulus features against irrelevant, distracting ones. (c) 2015 APA, all rights reserved).